PDA

View Full Version : ARDI Flies Wildfire.



spydmobile
07-19-2011, 03:36 PM
This is not an aeroquad, but this is where many of my aeroquads are destined to go... For those who apprciate a bit of extreme quadcoptering, here is an excerpt from a 30 minute mission I did at work this summer, I wanted to share this very special aspect of our favorite pasttime, since I get to do it and get paid, however we are posting since quad footage like this is rarely seen:

7 Minute Summary: http://vimeo.com/26611523

Full 30 Minute mission: Coming Soon

Project FireEye 2011 (Government of the Northwest Territories and Western Willow Ventures Inc.)
Mission: Wildfire ZF12 - 140km north of Fort Providence NWT, Canada

Priority mission objectives (Operational): Using Micro-UAV and a ground camera, document the burnout action on fire ZF12, capture any possible extreme fire behaviour from altitude and afar.

Secondary mission objectives (Fire Sciences): Prove the ability of MUAV Platform to be flown, capture HD video documentation, and survive intact, a close up encounter with a wildland fire (intensity class 4)

Micro-UAV (Flying Western Willow Ventures Inc's A.R.D.i - Aerial Research Drone 1 - Flying a 30 cm Mk Quad, FC only, with 2.4 FASST

Camera 1: Gopro HD Hero R3 720p/60fps - Airborne (ARDi)

Camera 2: Panasonic TM700 1080p/60fps - Ground Tripod mount

The flying conditions were difficult due to the shifting winds, before the burnout action, once the two sides of the fire met in the middle, it was like flying blind in a hurricane. Myself (The Pilot) and the Mission Commander (Wearing FPV Goggles in chair beside me) both suffered minor and superficial burns to skin and clothing from flying debris during the mission, When the Commander evacuated the mission base (my SUV) I remained and I flew until the battery was done. Burning conditions were not ideal, but sufficed. temps were high, humiditys low, winds were steady.

This form of documentation is impossible in conventional aircraft becuase they alter the behaviour of the fire, while MUAV does not as evidenced by this footage.

I have left the timecode onscreen so that viewers understand the time elapsed during the full mission, which lasts about 30 minutes, while this video cuts out a great deal of the mission. A link has been provided to the full 30 minutes for those who are intrested.

Pilot notes:

~5:30 into the mission: Downdraft turns my low battery decent into a powerful quadcopter trampoline jump. Thank godness for the landing gear/ pontoon rig. my first indicator of the chaotic nature of the winds around a wildfire, even in its infancy.

~15:20 into the mission: Downdrafting, updrafting, and winds in all directions impede shooting, the quad is thrown around like a plastic bag in the wind.

~16:30 into the mission: A flare up, begins intensifying the fire on the east (Right) side of the area.

~17:35 into the mission: The west (left) side of the area is now burning as well...
~18:16 into the mission: The right side nears full intensity and my supervisor yells "Go ARDi Go!"
~19:24 into the mission: An aerosol can or other combustible refuse can be heard exploding in the forest.
~20:48 into the mission: The left side nears full intensity.
~21:20 into the mission: Last battery change before things go full tilt.
~22:00 into the mission: Both sides of the fire, begin to come together and the blaze intensifies into a wall of fire.
~22:25 into the mission: I fly ARDi into the thermal column of the fire, and a wild ride into flying debris begins
~23:00 into the mission: wind shifts, and smoke column and firey debris rains down on the crew and ARDi - hot debris, sticks to the plastic gopro lens
~23:20 into the mission: both fires join, and ARDi is pummeled with debirs, I bring it down to about 10 feet and fly it into the main fire.
~24:17 into the mission: Mission commander, gets covered in burning debris and evacuates the position.
~25:30 into the mission: A small rabbit, escapes the blaze to safety, I follow its path with ARDi

chris1seto
07-19-2011, 03:47 PM
Were you able to fly for 30 minutes on one battery? Also, specs of the aircraft?

spydmobile
07-19-2011, 04:29 PM
Hi, no I flew 4 batteries on this mission.
Specs, I posted above that it is a 30 cm Mk Quad, FC only, with 2.4 FASST but for more specs here ya go:
BL-ESC: MK BL-Ctrl V1.2
Baro: MPX 4115A
Flight Controller: Mikrokopter ME V 2.0
Frame: Mikrokopter MK-30
Frame Construction: Aluminum & Fiberglass
GPS: MK-LEA4H GPS (u-blox)
Motors: Roxxy 2827-35 - black Brushless Motor
Platform Format: Quadcopter Plus
Power: Turnigy 4S 40C 2650 mAh Lipo
Props: EPP 10x45
rx: Futuba R6017FS

Honk
07-19-2011, 04:55 PM
Amazing footage, especially for someone who has never seen a wildfire!

Do they happen often enough for you guys to get practice/routine on this? How do you manage to set everything up in time?

Did you have any camera servo on the GoPro? Didn't seem like it. While it was totally amazing and great shot, I think it would have been interesting if you also had flown higher above it and pointed the camera almost straight down into the fire, but that might not have been possible due to heat/wind?

ridgebackred
07-19-2011, 05:35 PM
Very lifesaving research spydmobile.

Were you flying line of sight?

Do you think a heavy lift octo could cope with those conditions better than a quad?

spydmobile
07-19-2011, 05:43 PM
Amazing footage, especially for someone who has never seen a wildfire!

Do they happen often enough for you guys to get practice/routine on this? How do you manage to set everything up in time?

Did you have any camera servo on the GoPro? Didn't seem like it. While it was totally amazing and great shot, I think it would have been interesting if you also had flown higher above it and pointed the camera almost straight down into the fire, but that might not have been possible due to heat/wind?

Thanks Honk!!
No they are rare near a road, and even more rare when I have all my gear, the territory I am responsible for is actually 3 x the size of sweden and we have 5 highways. we were 40kms from the fire doing science experiments and perparing to burn our forest plots when we were told to shut down becuase of the nearby fire, so we worked with the Operations folks and moved our gear to the real fire.

ARDI has had many diffent camera mounts including automatic and manual gimbals, but I find that a solid mounted gopro using my custom mounting system provide a nearly shake free video and I am able to produce nicer videos that when we connect servos. Meanwhile removing them provided extra payload cpacity for loggers and sensors.

The full 30 minute version of the video has lots more fire stuff, the 7 min version is all about the UAV, while the full 30min version is all about the fire. I will let you know when I have that one online.

Last year in the fall I flew ardi directly above some small slash fires in the national park and found that as soon as you get in the column, you get an instant free ride up to 500 ft ;-) very hard to see and control like that!

Lastly, FPV is impossible in these situations becuase of the 180+ degree blind spot, and the risk to aircraft and people, so no FPV with fire :(

Franco

spydmobile
07-19-2011, 05:47 PM
Very lifesaving research spydmobile.

Were you flying line of sight?

Do you think a heavy lift octo could cope with those conditions better than a quad?

Hi, Yes, FPV is impossible in these situations becuase of the 180+ degree blind spot, and the risk to aircraft and people, so no FPV with fire :(

I have 13 UAVs and I think that smaller heavier is better, my radial Octo (1m motor to motor) is dramatically affected by wind change while my kinjal Aeroquad ( FC=APM) with two batteries and a gopro hangs nice in the wind and is almost unaffected affected by wind. There is still lots of research left to do. the lightweight pool noodle floats cause a lot of interference from wind. I have A DEX quad coming soon and that may prove to be very different aerodynamically speaking.....
Franco

Honk
07-20-2011, 07:51 AM
Ah! Too much knowledge to take in! You seem to know exactly what I wanna know... Just have a hard time taking it all in...


Last year in the fall I flew ardi directly above some small slash fires in the national park and found that as soon as you get in the column, you get an instant free ride up to 500 ft ;-) very hard to see and control like that!

Lastly, FPV is impossible in these situations becuase of the 180+ degree blind spot, and the risk to aircraft and people, so no FPV with fire

So, flying above fire (hot air) gets it easier for the quad to lift? I might be tired, but I thought it was the inverse? That cooler air made the air denser thus easier for the props grabbing onto the air so to speak?

If the field of view isn't enough, wouldn't a yaw servo (attached to the cam) controlled by a IMU/compass on your head solve the whole thing? Then you could just look in any direction you wanted (well limited to servo travel, might be able to have some transmission on it though).


You also mention that you have went from servo's to fixed position... I have a very hard time realizing that it would be better without any servo's, I like the ability to change the camera angle in real time, but since you seem a heck of a lot more experienced, now I'm having the hardest time deciding! When do I actually want camera servo's...

And oh oh oh, was that last statement of yours screwing with my summer brain! Smaller = more stable in wind! Does not compile for me! I tried to make some video's in central Stockholm some days ago with a Snapquad and 10" props, and the wind just grabbed it (like a plastic bag as you described earlier) any time it was present. Flying near the waters (Mälaren) was just horrible, going into alleys/lanes was fine until the wind changed direction and it was just sucked straight into the alley without control. Flying in the garden last day the wind made it wobble around like a little ball...

While my V-shaped octo seemed to cope with small wind gusts pretty well (although making some slow swings afterwards). Also with 10" props, obviously twice as many.

So, might my problem actually be that I have too big props on the smaller quad? And will the "smaller copes with wind better than larger" always hold true given you have the right props?



And yeah, I saw that DEX quad the other day too (strange coincidence, I rarely check RC groups because there's just too much :) ) and it looked awesome! Apparently great for being a boat too! :P

M@d_Hunter
07-20-2011, 11:05 AM
So, flying above fire (hot air) gets it easier for the quad to lift? I might be tired, but I thought it was the inverse? That cooler air made the air denser thus easier for the props grabbing onto the air so to speak?

You forgot that very hot air with a lot of cold air around will create a strong ascending wind.

Honk
07-20-2011, 12:18 PM
Wow, yeah I realize that when you say it, but no, I didn't know it! Nice... So cool air gets like sucked in and forms a pile of doom? Gotta try it :) Might be the ticket to longer flight times!

BTW, for a fire adapted quad/multirotor, what would be the criteria's? I'm thinking a quad that can go close to fires both outdoors and indoors... Maybe IR/thermal camera (flir?) and such too?

ridgebackred
07-20-2011, 01:52 PM
Honk,
Some of the data that spydmobile is collecting is probably the maximum thermal updraft velocity of the air around wildfires. (We do have all the sensors needed!)
Logging the motor power setting at the same time as the sensor data should allow a normalization of the data to compensate for motor thrust.

The octo, quad observation about controllability had me confused until I think about gliders vs short wing planes. The prop on each motor not only provides thrust, it is adding to the effective "wing area" of the copter. The more prop disk area, the more the copter is affected by vertical movement of the surrounding air mass.

I will take spydmobile's word that a heavy (ish) quad with more powerful motors than normal on a sport machine is near optimum for wildfire research.

Further optimizations that I can think of are carbon fiber props for least weight and fastest response and reducing the flat plate area of the frame for aerodynamics.

Honk
07-20-2011, 02:04 PM
Hmm yeah well, in the future (timeline unknown) I think we'll have the accelerometer and barometer working great together when we start using the proper accel cal, so yeah, then vertical velocity will be easy to measure. But I don't think that the crafts' vertical velocity will entirely reflect the true air velocity, so for that measurement I figure one would want a airspeed sensor instead.... Might be useful for craft controls too though (detecting translational flight and help leans? or just help when doing long turns...).


Yes, I have thought about that too. When trying 12" props before they acted like huge sails just grabbed by the wind really easily... But then I must ask this counter question: wouldn't just smaller props be the solution to ANY mutlirotor which you want to fly good in the wind? I definitely gotta get hold of some smaller props and try on both my quad and octo. My hypothesis is that the octo with smaller props will still do better than a small quad with the same change (going from 10 to 8" props perhaps).

spydmobile
07-20-2011, 02:16 PM
Honk,
Some of the data that spydmobile is collecting is probably the maximum thermal updraft velocity of the air around wildfires. (We do have all the sensors needed!)
Logging the motor power setting at the same time as the sensor data should allow a normalization of the data to compensate for motor thrust.

The octo, quad observation about controllability had me confused until I think about gliders vs short wing planes. The prop on each motor not only provides thrust, it is adding to the effective "wing area" of the copter. The more prop disk area, the more the copter is affected by vertical movement of the surrounding air mass.

I will take spydmobile's word that a heavy (ish) quad with more powerful motors than normal on a sport machine is near optimum for wildfire research.

Further optimizations that I can think of are carbon fiber props for least weight and fastest response and reducing the flat plate area of the frame for aerodynamics.

Bingo - thats this ticket, the bigger props, and more area = more vulnerability, so I find that when I pile all the same gear on a smaller quad I get more stability (in principle) but I also get more invincible to the chaotic wind forces generated by the indrafting and column development, and contrary to popular belief among the other scientists, I did not have to be near the column for this to happen, just in the vicinity. the simple act of changing the 2560 mha battery out for the 5000 mha one instantly made ARDI respond slower, but inherently more stable in the column. I now practice flying near home on the edge of a couple of land formations (like cliffs) where diurnal heating and cooling cause kadiabatic type wind changes, while this is not a wildfire smoke column, nor wildfire heat column development it is chaotic to some extent and it helps me tune the quad and my hands to perform for fire.

The data currently being collected for altitude is feeding off the eagletree GPS only, as I have had issues with the MK GPS so I do not use it.This means that the altitude data is not high enough resolution nor accurate enough for that part of the science, and there is thermal imaging / software that analyzes video and determines updrafting velocity. Also, at 5000$ ARDi is NOT the quad I want to keep using in wildfire, I want to build AeroQuads that do this work, But have been using APMs with AQ2.4 instead for lack of GPS etc. Lastly, the real issue here is to prove that this can be done, once I have won over the fire community (And we are well on our way) THe science will change, but I am leaning to operational use rather than pure science. and I am succeding with the 1-2 punch approach, This video is the first punch we hit them with then - Rotary Wing Aircraft on a fire start at 3000$ /hour + fuel. You risk a million dollar aircraft and human life. MUAV on wildfire is a no brainer. The main hurdle is law and federal aviation regulations, but in this line of work, that is not an issue becuase there is the NOTAM. When on a fire, the Incident Commander has absolute control and authority of the airspace, that means I fly regardless of RC rules, and federal aviation rules become secondary to the NOTAM (which establishes the new restricted airspace around the fire which no other aircraft may enter, but I get to fly in) and say again, I love my job!
:cool:

ridgebackred
07-20-2011, 02:23 PM
I tried smaller and smaller props with higher and higher pitch on my tricopter (probably reducing efficiency all along the way).
Everything was great with higher frequency vibration (my gyro's like that) and more stability until I ran into prop stall at hover.
The addition or reduction of just a few grams of payload make a huge difference when using small props with high pitch at high rotation speed.
The knee curve for prop stall seems very sharp.

Frame aerodynamics is tougher to optimize. Copter flight (and gusts in hover) occur in all directions.
An airfoil shape in the forward direction of flight isn't really possible.
Dimpled tape (available from full size sailplane supply houses) wrapped around circular cross section frame arms may be the best.

spydmobile
07-20-2011, 05:35 PM
Nice this supports what I am seeing from units like this one: http://shrediquette.blogspot.com/p/shrediquette-bolt.html
Incredible stability speed and performance!

ridgebackred
07-20-2011, 07:29 PM
If a digital servo could respond *much* faster than the fixed pitch motor speed response for controlling thrust, these may be worth the added weight and complexity.
The processing horsepower for the mixing table is certainly available.
They are in the quadcopter power range and off-the-shelf.


http://www.e-fliterc.com/Products/ProductGallery.aspx?ProdID=EFLPVPP100

M@d_Hunter
07-21-2011, 11:26 AM
Hmm yeah well, in the future (timeline unknown) I think we'll have the accelerometer and barometer working great together when we start using the proper accel cal, so yeah, then vertical velocity will be easy to measure. But I don't think that the crafts' vertical velocity will entirely reflect the true air velocity, so for that measurement I figure one would want a airspeed sensor instead.... Might be useful for craft controls too though (detecting translational flight and help leans? or just help when doing long turns...).

To measure air velocity, I only know pitot tube. It seems to exist for RC planes but I'm not sure that it's very accurate. And with all our propeller, there will be a lot of "interference" due to the air movement.

Mikro
07-21-2011, 02:31 PM
Hey guys, lurking in the background here... but am so impressed with your work Spydmobile! Thanks for sharing! I'm a bit surprised that smaller/heavier has given you better stability, as the traditional line of thought is that larger is more stable (but usually larger means more heavier). My Cyclone frame prototypes are 24" motor to motor, so I'll give it a go and try a smaller configuration to see what happens with my setup.

In this situation, does GPS waypoint type flight help? (As you indicate that FPV flight is not desirable here).